I had originally done this (ext2 rather than ext3) several years ago because of some unrecoverable corruption I had on an ext3 filesystem which I feel I could have repaired (or wouldn't have happened) on an ext2 type. The Wikipedia page for ext3 also makes light mention of potential risks.
These days (ext3 was very new at the earlier time) I think ext3 is probably a better choice and more stable... certainly faster at recovery on 26GB than ext2 will be, so I'm going to covert the partition from ext2 to ext3. (As I write this ext4 is currently marked as "unstable", but will eventually overcome some of the negative issues that ext3 currently has).
Anyway, after spending some time on Google and doing some reading, I've come to the conclusion that I'll use the following steps:
- Code: Select all
umount /dev/sda9 # better to make sure it's unmounted
fsck -f /dev/sda9 # better to make sure it's currently clean
tune2fs -c 0 -i 30 -j /dev/sda9 # -c = fsck for mount count off, -i = we'll fsck every 30 days, -j = add journal
sed -i '/sda9/ s/ext2/ext3/' /etc/fstab # change fstab to now use ext3 for /dev/sda9
mount /dev/sda9 # remount the partition
After doing this on my 26GB partition, I noticed that it now shows about 130MB less available disk space (used by the journal, no doubt).
One of the better alternative pages I found is here:
http://www.math.ucla.edu/~jimc/document ... 2ext3.html
From this point I hope to be seeing significant speed enhancement during fsck usage on that large partition in the future.
Perhaps also a little more reliability, but I'll certainly be upgrading to ext4 after it becomes stable.
david
